Methods for allocating transmission bandwidths of a network

Abstract

Methods for allocating transmission bandwidths of a network are adapted to a network including an office terminal and peripheral terminals. The peripheral terminals communicate with the office terminal by time division multiplexing during a sequence of transmitting cycles. The method includes receiving requested bandwidths from uploading messages delivered from the peripheral terminals, arranging an uploading order of the peripheral terminals based on the uploading messages to obtain a transmitting sequence, adjusting the uploading order of each of the peripheral terminals in the transmitting sequence based on a size of the requested bandwidth to obtain a modified transmitting sequence, and arranging a modified uploading order of the peripheral terminals based on the modified transmitting sequence. Therefore, the transmission bandwidth allocation is fairer, and delay is reduced. Upstream order of each terminal is transferred based on its requested bandwidth, thereby effectively reducing the average delay.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a continuation of non-provisional U.S. application Ser. No. 11/363,941, filed on Mar. 1, 2006, which claims priority under 35 U.S.C. §119(a) on Patent Application No(s). 094132107 filed in Taiwan, R.O.C. on Sep. 16, 2005, the entire contents of which are hereby incorporated by reference. U.S. application Ser. No. 11/363,941 issued as U.S. Pat. No. 7,577,162 on Aug. 18, 2009.

TECHNICAL FIELD

The invention relates to a method for uploading data, and in particular to a method for allocating transmission bandwidths of a network.

BACKGROUND

For a long time now bandwidth allocation has been an important subject in designing a network system. Taking a passive optical network (PON) as an example, multiple optical network units (ONU) are disposed at a corresponding number of offices or houses, using passive devices to couple to a single optical line terminal (OLT). In other words, an optical line terminal located at one end connects to an optical couple device, which is near a terminal (client side), by an optical fiber, and then goes to the optical network unit. Here, data can be transmitted to the optical network unit by the broadcasting of the optical line terminal, which is called a downloading process. On the other hand, as shown in FIG. 1, an uploading process means an optical network unit performs a time division multiplexing (TDM) process to transmit the data to the optical line terminal. While uploading data, the uploading bandwidth is shared by all optical network units, therefore bandwidth allocation directly affects transmission speed when uploading data and the efficiency of the bandwidth. However, the current method of allocating bandwidth is not provide the required properties such as low transmission delay, high bandwidth efficiency and fairness of bandwidth allocation.

Traditionally, each optical network unit is allocated at the same portion of a whole bandwidth (i.e. in one time division multiplexing channel) and the transmissions of those optical network units are synchronized to prevent collision (e.g. two or more optical network units have partially overlapping transmission). For instance, in the related art, N pieces of optical network units are separately assigned to a time slot and adapted to it. Every optical network unit can transmit any number of data package as shown in FIG. 2. Here, if some package cannot complete the transmission in the current time slot, this package must be retained and wait until the next time slot to transmit. Although there will be no collisions and package separations occurring in this method, this time slot allocation method with circulation fixing is not handle a situation like bursting net flow.

Therefore, a dynamic bandwidth allocating method has been provided. This method is able to reduce the corresponding size of time slots when no data package is transmitted, and give the remainder of the bandwidth to other optical network units to use. However, according to this method, in order to receive time slot assignments precisely, the optical line terminal must acknowledge how many bits of data packages are waiting in every optical network unit before the assignment. Here, every optical network unit transmits a specific message to inform the optical network terminal how many bits of data packages are sent before the data transmission. Then the optical line terminal is able to estimate and allocate the bandwidth to the optical network unit which is going to transmit a data package, inform the optical network unit about the transmittable bandwidth, and start data package transmission. During the uploading process, the optical line terminal monitors the transmission by the optical network unit in a proper order for arranging the transmission timing of the next optical network unit, so that it can receive the next transmitting data package from the next optical network unit after the last data transmission.

In the dynamic bandwidth allocation structure, lots of algorithms for the dynamic bandwidth allocation are provided. For example, U.S. patent application 20030048805 A1 provides an algorithm for the dynamic bandwidth allocation, shown in FIG. 3. In this figure, this method sets up an assured bandwidth B_min_j for every optical network unit, where j is the number of the terminals with the requested bandwidth, or the number of optical network units with the requested bandwidth. In other words, if a usable bandwidth of the nth cycle is B_ref, the usable bandwidth of the optical network unit in the nth cycle will be allocated based on the proportion of the assured bandwidth B_min_j. After that, in the nth cycle, an ideal usable bandwidth B_ideal_j,n for every optical network unit with requested bandwidth can be obtained by the allocated usable bandwidth subtracting the compensation value B_add_j,n−1. Next, bandwidth allocation proceeds. When the ideal bandwidth B_ideal_j,n of the optical network unit is larger than 0, the allocated transmittable bandwidth B_temp_j,n of the optical network unit is the bandwidth Q_j,n requested; otherwise the allocated transmittable bandwidth B_temp_j,n is 0. After the bandwidth allocation is completed and the compensation value B_add_j,n is recalculated, the allocation of the uploading bandwidth for the optical network unit is completed.

The compensation value is the sum of overspent bandwidths accumulated before this cycle of each optical network unit. When the compensation value is larger than 0, there is an overspent bandwidth before this cycle. On the other hand, when the compensation value is smaller than 0, there is a remaining bandwidth before this cycle.

Under this structure, although a fairness of bandwidth allocation can be obtained, problems remain. If the optical network unit doesn't use many bandwidths for a long time, the remaining bandwidth will be largely accumulated. And when this optical network unit suddenly produces a large amount of bandwidth requests, a large amount of bandwidth will be exhausted, leading to a longer transmitting delay to other optical network units. On the other hand, if the optical network unit maintains a large amount of bandwidth requests for a long time, a usable ideal bandwidth can hardly be stable, leading to a large jittering of the transmitting delay.

In addition, other related documents show another dynamic bandwidth allocation algorithm. See Chadi M. Assi, Yinghua Ye, Sudhir Dixit, and Mohamed A. Ali, “Dynamic Bandwidth Allocation for Quality-of-Service Over Ethernet PONs, “IEEE Journal on Selected Areas in Communications, Vol. 21, No. 9, November 2003, pp. 1467-1477. This method allocated the usable bandwidth to each optical network unit based on the assured bandwidth and the requested bandwidth in this cycle, and then allocated the overspent bandwidth (i.e. the remaining usable bandwidth) based on the requested bandwidth, which is represented by the following formula.

$\begin{array}{cc}{\mathrm{B\_grant}}_j=\{\begin{array}{cc}{R}_j,& \mathrm{if}{R}_j\le {\mathrm{B\_min}}_j\\ {\mathrm{B\_min}}_j+{\mathrm{B\_excess}}_j,& \mathrm{if}{R}_j>{\mathrm{B\_min}}_j\end{array}& \left(1\right)\\ {\mathrm{B\_min}}_j=\frac{\left(T_{\mathrm{cycle}}-N\times T_g\right)\times r}{8}\times w_j& \left(2\right)\\ {\mathrm{B\_excess}}_j=\frac{\mathrm{B\_left}\times R_j}{\sum _{k\in K}{R}_k}& \left(3\right)\end{array}$

Here, the formula (1) is the basic method for the major bandwidth allocation, where Rj is the requested bandwidth for every optical network unit; B_min_j is the assured bandwidth for every optical network unit; B_excess_j is the excess bandwidth reallocated to the optical network unit with the remaining bandwidth; and B_grant_j is the bandwidth (i.e. transmittable bandwidth) the optical network unit actually received. According to the formula (1), when the requested bandwidth is smaller or equal to the assured bandwidth, the optical network unit receives the bandwidth requested; otherwise the optical network unit receives the excess bandwidth in addition to the assured bandwidth. Here, the assured bandwidth can be calculated by the formula (2), where T_cycle is the cycle time; N is the number of the optical network units; T_g is the switching time of the optical network unit; r is the transmitting rate; and w_j is the assured bandwidth weight value of the optical network unit, which is determined by the costumer contract. Furthermore, the excess bandwidth can be calculated by the formula (3), where B_left is the remaining bandwidth of the usable bandwidth in this cycle after the allocation of the assured bandwidth, and K is a class of the massive loading (i.e. the requested bandwidth is larger than the assured bandwidth) optical network units, which is represented by K={R_j>B_min_j}.

Although this framework can use the remaining bandwidth effectively, it allocates the remaining bandwidth based on the proportion of the requested bandwidth of the optical network unit, which is not fair. So costumers who only buy less bandwidth may have large amounts of transmittable bandwidth to use just because they request it.

Therefore, another method was provided to use the flow prediction process to previously allocate the spare bandwidth in order to reduce the holding time of the high priority data, which is accomplished by the following formulas.
R_j=(H_j+E_wait_j(n))+M_j+L_j  (4)
E_wait_j(n)=A_wait_j(n−1)  (5)

According to the formula (4), the requested bandwidth of the optical network unit consists of requested bandwidths with high (H_j), medium (M_j), and low (L_j) priorities, where a flow prediction value (E_wait_j(n)) is added to the high priority part and n is the number of the cycles. From the formula (5), A_wait_j(n−1) represents the amount of data actually reaching high priority in the holding period of the n−1th cycle. This shows that the flow prediction value (E_wait_j(n)) is the requested bandwidths for high priority reaching the holding period of the last cycle.

However, although this process may reduce the averaged transmission delay of the high priority data, it increases the average transmission delay of other priority data and the inaccuracy of the flow prediction reduces the efficiency of the bandwidth.

Thus, how to effectively provide low delay transmission, high bandwidth efficiency and fairness of bandwidth allocation becomes very important research in this bandwidth allocation field.

SUMMARY

According to an exemplary embodiment, a method of allocating network bandwidth is further provided where the network includes an office terminal and multiple terminals connecting to the office terminal. This method includes: receiving requested bandwidths from uploading messages to obtain a transmitting sequence, sorting an uploading order of the peripheral terminals based on the uploading messages to obtain a transmitting sequence, adjusting the uploading order of each of the peripheral terminals in the transmitting sequence based on a size of the requested bandwidth to obtain a modified transmitting sequence, and sorting a modified uploading order of the peripheral terminals based on the modified transmitting sequence.

Here, by repeatedly performing the step of adjusting every terminal's uploading orders, every terminal's uploading order is sorted to become a better sequence according to the amount of data uploaded. However, in order to prevent the uploading order of every terminal from changing too much, a predetermined value can be set up first. Also, after receiving the modified transmitting sequence, this modified transmitting sequence is checked to determine whether it is the same as the original transmitting sequence or not. If they are not the same, the changing number is added up and then confirmed as to whether it reaches the predetermined value. If the cumulated changing number doesn't reach the predetermined value, this process will go back to the adjusting step to base on the requested bandwidth to adjust the transmitting sequence again; on the other hand if the cumulated changing number does reach the predetermined number, the adjustment for the transmitting sequence will stop and the office terminal will use the last obtained modified transmitting sequence to be the order for determining the terminals to upload.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention will become more fully understood from the detailed description given below, which is for illustration only and thus is not limitative of the invention, wherein:

FIG. 1 is a diagram showing a basic structure of a conventional network;

FIG. 2 is a diagram showing the method of uploading data in the conventional network of FIG. 1;

FIG. 3 is a diagram showing a conventional method of allocating the transmitting bandwidth;

FIG. 4 is a flow chart showing an exemplary embodiment of a method for allocating a bandwidth of a network;

FIG. 5 shows an embodiment of the relationship between the weight value of FIG. 4 and the loading extent of the network;

FIG. 6 shows another embodiment of the relationship between the weight value of FIG. 4 and the loading extent of the network;

FIG. 7 shows still another embodiment of the relationship between the weight value of FIG. 4 and the loading extent of the network;

FIG. 8A is a flow chart showing one exemplary embodiment of a method for allocating a bandwidth of a network;

FIG. 8B is a flow chart showing another exemplary embodiment of a method for allocating a bandwidth of a network;

FIG. 9A is a flow chart showing another exemplary embodiment of a method for allocating a bandwidth of a network;

FIG. 9B is a flow chart showing another exemplary embodiment of a method for allocating a bandwidth of a network;

FIG. 10 is a flow chart showing another exemplary embodiment of a method for allocating a bandwidth of a network;

FIG. 11 is a detailed flow chart showing an exemplary embodiment of step 220 in FIG. 10;

FIG. 12A is a detailed flow chart showing an exemplary embodiment of step 240 in FIG. 10;

FIG. 12B is a detailed flow chart showing another exemplary embodiment of step 240 in FIG. 10;

FIG. 12C is a detailed flow chart showing still another exemplary embodiment of step 240 in FIG. 10;

FIG. 13 is a detailed flow chart showing an exemplary embodiment of step 245 in FIGS. 12A to 12C;

FIG. 14 is a detailed flow chart showing an exemplary embodiment of step 250 in FIG. 10;

FIG. 15A is a flow chart showing another embodiment of a method for allocating a bandwidth of a network according to the present invention;

FIG. 15B is a flow chart showing another exemplary embodiment of a method for allocating a bandwidth of a network;

FIG. 16A is a detailed flow chart showing one exemplary embodiment of step 260 in FIG. 15A;

FIG. 16B is a detailed flow chart showing another exemplary embodiment of step 260 in FIG. 15B;

FIG. 17 is a flow chart showing another exemplary embodiment of a method for allocating a bandwidth of a network;

FIG. 18A is a detailed flow chart showing one exemplary embodiment of step 330 in FIG. 17;

FIG. 18B is a detailed flow chart showing another exemplary embodiment of step 330 in FIG. 17;

FIG. 18C is a detailed flow chart showing still another embodiment of step 330 in FIG. 17; and

FIG. 19 is a flow chart showing still another exemplary embodiment of a method for allocating a bandwidth of a network.

DETAILED DESCRIPTION

In the exemplary embodiments below, the disclosed embodiments can apply to a network that includes an office terminal and multiple terminals connecting to the office terminal. Hardware that applies in the related art can also be disposed at the office terminal and the multiple terminals to carry out this network if appropriate.

For example, the applied network can be a passive optical network (PON), so there can be an optical line terminal (OLT) at the office terminal and an optical network unit (ONU) at each terminal. The optical network unit and the optical line terminal can separately have their own central processing unit (CPU) to control an operation of the media access control (MAC) logic circuit. Among them, every MAC logic circuit can be included in one single integrated circuit (IC), such as the MPC860TZP50 interface, the RS232 interface and the 10BaseT interface from Motorola. Besides, the optical network unit and the optical line terminal also can include a network processor chip such as the IXP1200 from Intel, the MXT-4000 series and the MXT-5000 series from Maker (Conexant), the Prism from Sitera and the nP3400 from MMC to perform the Ethernet network's packaging process. Here, the network processor chip also can include a MAC chip such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA) to provide the access to the network. Also, these optical network units and the optical line terminal can further include a memory (ex. read-only memory, ROM) or a random-access memory (RAM) or can use an optical transponder to perform two way transmissions by an optical fiber. Although the network mentioned in this specification can use any kind of optical transponders, one of methods can include using a transponder that is capable of using in an integrated circuit and transmitting and receiving with 1.3 μm wave meter and 1.55 μm wave meter respectively (ex. a planar light wave circuit (PLC)), and using a forward feedback circuit (ex. a ROM) to work not instantaneously with a bursting first bit to work under a transmission speed of 1.25 Gbps. However, the hardware used in the optical network unit or the optical line terminal is not the key component, which means the invention can use any known hardware adapted to the disclosed embodiments.

The processes mentioned below are generally performed by the above MAC chip, including the access to the network. It also can be performed by software which is executed and loaded by a CPU. The CPU is separated from but coupled to the MAC chip of the network.

In order to accomplish a network transmission bandwidth allocation method of the disclosure, there are three exemplary embodiments of bandwidth allocation methods provided respectively for adjusting the transmitting sequence based on the amount of uploading data, predicting the bandwidth allocation ratio based on the extent of network loading and allocating the transmitting bandwidth for conforming to the fairness of bandwidth allocation.

In an embodiment, the allocation for the network transmitting bandwidth is improved by the estimate of the predicting bandwidth. Here, the predicting bandwidth can be estimated by the following formula.
R_j=Q_j+E_j×W(L)  (6)

In this formula, R_j is the estimated requested bandwidth of the terminal, Q_j is the data amount waiting for transmission in the terminal, E_j is the data amount expected to reach the terminal in a holding time, L is the extent of network loading in this transmitting cycle and W(L) is the weight value changed with the extent of the network loading. Thus, according to the formula (6), the allocation ratio of the predicting bandwidth will be adjusted based on the extent of the network loading.

J is the number of service terminals requesting bandwidth as well as the number of the terminal of request bandwidth. Therefore, in a network, the office terminal can base on the estimated requested bandwidth to estimate the requested bandwidth of the terminal one by one by the formula (6) and further determine the amount of data (i.e. transmitting bandwidth) can be upload for uploading messages sent by the corresponding terminal.

Please refer to FIG. 4. In a transmitting cycle, first receive the transmittable bandwidth and the predicting bandwidth of the terminal (step 110); then adjust the predicting bandwidth based on a weight value (step 120); next estimate the terminal's requested bandwidth by the formula (6) and add the transmittable bandwidth to the modified predicting bandwidth to obtain the final requested bandwidth (step 130).

Here, in order to prevent wasting the bandwidth, which is contributed by the prediction error in predicting the bandwidth E_j, a larger weight value W(L) can be introduced when the loading of the network is low; on the other hand, a smaller weight value W(L) can be introduced when the loading of the network is heavy. That is, the weight value W(L) decreases with the increase of the extent of network loading L. The relationship between W(L) and L are shown in FIGS. 5, 6 and 7. As a result, the bandwidth will not be wasted when the network loading is heavy, and the transmission delay can be shortened when the network loading is low by the prediction of the bandwidth. Next, the predicting bandwidth can be a transmitting bandwidth in the last transmitting cycle for every terminal, which is the data uploading amount in the last transmitting cycle.

In other words, the weight value in step 120 can be obtained by calculating the loading extent of the network (step 140) first and then by using the loading of extent as a basis (step 150). This weight value can be obtained before step 110 or until before step 120, as shown in FIGS. 8A and 8B.

Practically, the above steps can be continuously repeated to estimate the requested bandwidth for terminals to which uploading messages are delivered, as shown in FIG. 9A. Or, receive the transmittable bandwidth and the predicting bandwidth for terminals to which uploading messages are delivered first, then followed by continuously repeating steps 120 to 130 to estimate the requested bandwidth for terminals to which uploading messages are delivered, as shown in FIG. 9B.

Please refer to FIG. 10. Obtain the requested bandwidths of at least one terminal to which uploading messages are delivered (step 210), then calculate the transmitting bandwidth allocated to the terminals requesting bandwidth based on the obtained requested bandwidth. Here, the transmitting bandwidth is the data amount allowable for the corresponding terminal uploading through the network. Every terminal has its useable bandwidth range, and the useable bandwidth range is between an assured bandwidth and a maximum bandwidth.

After obtaining the requested bandwidth, the initial bandwidth allocation will first proceed. This cycle's usable bandwidth will be allocated to every terminal based on the requested bandwidth and the assured bandwidth of each terminal, which means allocating every terminal's transmitting bandwidth one by one (step 220). Compare every terminal's requested bandwidth and assured bandwidth (step 221). The comparison result is the basis for distributing a transmitting bandwidth that conforms either to the requested bandwidth or to the assured bandwidth to the corresponding terminal. If the requested bandwidth is smaller or equal to the assured bandwidth, the transmitting bandwidth allocated to this terminal will be the requested bandwidth for this terminal (step 223). If the requested terminal is larger than the assured bandwidth, the transmitting bandwidth allocated to this terminal will be the assured bandwidth for this terminal (step 225), as shown in FIG. 11. Also, the above steps (the step 221 and the step 223 or the step 221 and the step 225) can be repeated to accomplish the initial allocation of the transmitting bandwidth for every terminal to which uploading messages are delivered.

When the requested bandwidth is larger than the assured bandwidth, an additional excess bandwidth will be allocated to this terminal in addition to the transmitting bandwidth that conforms to the assured bandwidth. Thus, after the initial bandwidth allocation (i.e. step 220), whether there is an unsatisfied requested bandwidth or not will be confirmed one by one (step 230). When there is an unsatisfied requested bandwidth, the remaining bandwidth will be further allocated to the terminal with the unsatisfied bandwidth, meaning an excess bandwidth will be further allocated to every terminal with unsatisfied requested bandwidth to obtain the reallocated transmitting bandwidth of every terminal (step 240), as shown in FIG. 10. In other words, step 230 is to compare the allocated transmitting bandwidth and the requested bandwidth, where if the requested bandwidth is larger than the allocated transmitting bandwidth, an unsatisfied requested bandwidth exists. The remaining bandwidth is the result by the usable bandwidth of the cycle subtracting the allocated transmitting bandwidth (i.e. the remaining bandwidth amount) after the initial bandwidth allocation.

That is, distribute the usable bandwidth of this cycle to every terminal based on the assured bandwidth and the requested bandwidth by the following formula first.

$\begin{array}{cc}{\mathrm{B\_grant}}_j=\{\begin{array}{cc}{R}_j,& \mathrm{if}{R}_j\le {\mathrm{B\_min}}_j\\ {\mathrm{B\_min}}_j+{\mathrm{B\_excess}}_j,& \mathrm{if}{R}_j>{\mathrm{B\_min}}_j\end{array}& \left(7\right)\end{array}$

Formula (7) is the basic method of allocating the transmitting bandwidth, where R_j is the requested bandwidth, B_min_j is the assured bandwidth, B_excess_j is the excess bandwidth obtained by the reallocation of the remaining bandwidth, and B_grant_j is the bandwidth actually allocated to the terminal (i.e. the transmitting bandwidth). In the formula (7), if the requested bandwidth R_j is smaller or equal to the assured bandwidth B_min_j, this terminal will get the bandwidth it requests (i.e. the requested bandwidth R_j); otherwise this terminal will get the assured bandwidth B_min_j plus the excess bandwidth B_excess_j.

Compared to the related art, regarding to the calculation of the excess bandwidth, one of the embodiment is to calculate the excess bandwidth based on the maximum bandwidth and the bandwidth compensation value of the terminal with unsatisfied requested bandwidth. Therefore in step 240, the calculation of the excess bandwidth will be based on the usable bandwidth of the network in this cycle and the allocated transmitting bandwidth (step 241). Then calculate every terminal's allocatable extra bandwidth one by one based on the remaining bandwidth, the maximum bandwidth and the bandwidth compensation value of every terminal (with unsatisfied requested bandwidth) (step 243). Finally, further distribute the excess bandwidth one by one based on the extra bandwidth and the unsatisfied bandwidth of every terminal for adjusting the transmitting bandwidth of terminals with unsatisfied requested bandwidth (step 245), as shown in FIG. 12A.

The number for the excess bandwidth allocation in the detailed description is only one. However practically, according to the disclosed embodiments, whether there is an unsatisfied requested bandwidth or not will be confirmed after the allocation (step 246). If there is an unsatisfied requested bandwidth, it will perform the excess bandwidth allocation once more by executing the above steps (i.e. the step 241, the step 243 and the step 245) as shown in FIG. 12B. Besides, in order to prevent the number of the excess bandwidth reallocation from being to many, the number of allocation will be accumulated after every time of excess bandwidth allocation (step 247). While whether there is any unsatisfied requested bandwidth or not is confirmed, whether the number of the reallocation reaches a predetermined value or not will also be confirmed (step 249). If there exists an unsatisfied requested bandwidth and the number of the reallocation doesn't reach a predetermined value, step 241, step 243 and step 245 will then proceed to perform the excess bandwidth allocation once more, as shown in FIG. 12C.

Generally, the allocation method for the excess bandwidth is represented by the following formulas.

$\begin{array}{cc}{\mathrm{B\_left}}_n={\mathrm{B\_total}}_n-\sum _{j=1}^N{\mathrm{B\_min}}_j& \left(8\right)\\ {\mathrm{B\_extra}}_k=\frac{{\mathrm{B\_max}}_k-{\mathrm{B\_add}}_k}{\sum \left({\mathrm{b\_max}}_k-{\mathrm{B\_add}}_k\right)}\times {\mathrm{B\_left}}_n,\mathrm{and}k\in K& \left(9\right)\\ {\mathrm{B\_excess}}_k=\{\begin{array}{cc}{\mathrm{R\_left}}_k,& \mathrm{if}{\mathrm{R\_left}}_k\le {\mathrm{B\_extra}}_k\\ {\mathrm{B\_extra}}_k,& \mathrm{if}{\mathrm{R\_left}}_k>{\mathrm{B\_extra}}_k,\end{array}\mathrm{and}k\in K& \left(10\right)\end{array}$

Here, formula (8) is for calculating the remaining bandwidth, where B_min_j is the transmitting bandwidth allocated to the terminal (i.e. the assured bandwidth), B_total is the usable bandwidth of the network in the nth transmitting cycle (n is a positive), B_left is the remaining usable bandwidth in this network at present (i.e. the remaining bandwidth after the initial allocation of the usable bandwidth of the network in this transmitting cycle). According to this formula (8), the remaining bandwidth is the result of the usable bandwidth of the network subtracting the sum of the transmitting bandwidth allocated to every terminal.

Formula (9) is for calculating the extra bandwidth, where B_max_k is the maximum bandwidth transmittable by the terminal, B_add_k is the sum of the overspent bandwidth of the terminal cumulated until this transmitting cycle (i.e. the bandwidth compensation value of the terminal), B_extra_k is the extra bandwidth presently allocatable to the terminal, and K is a class of terminals with unsatisfied requested bandwidth (i.e. when the requested bandwidth R_j is larger than the assured bandwidth B_min_j), where K={R_j>B_min_j}. According to the formula (9), obtain a ratio of remaining bandwidth allocatable to each terminal by the maximum bandwidth B_max_k and the bandwidth compensation value B_add_k of every terminal first, then base on the ratio and the remaining bandwidth B_left_n to calculate the extra bandwidth B_extra_k allocatable to the terminal. The remaining bandwidth here is obtained by the formula (8). K and k are both positives.

Formula (10) is for allocating the excess bandwidth, where R_left_j is the remaining requested bandwidth by the requested bandwidth R_j subtracting the allocated transmitting bandwidth (i.e. the assured bandwidth B_min_j), which is R_left_j=R_j−B_min_j. According to the formula (10), if the remaining requested bandwidth R_left_j is smaller or equal to the extra bandwidth B_extra_j obtained from the formula (9), an excess bandwidth B_excess_j corresponding to the remaining requested bandwidth R_left_j will be further allocated; otherwise an excess bandwidth B_excess_j corresponding to the extra bandwidth B_extra_j will be further allocated. Here, every terminal's maximum bandwidth B_max_j and assured bandwidth B_min_j can be determined by the costumer contract.

In step 245 calculate the remaining requested bandwidth based on the requested bandwidth and the allocated transmitting bandwidth first (step 2451), then followed compare the remaining requested bandwidth and the extra bandwidth (step 2453). If the remaining requested bandwidth is smaller or equal to the extra bandwidth, an excess bandwidth corresponding to the remaining requested bandwidth will be further allocated to this terminal (step 2455); otherwise an excess bandwidth corresponding to the extra bandwidth will be further allocated to this terminal (step 2457) as shown in FIG. 13. Also, by repeatedly performing the above steps (i.e. steps 2451, 2453 and 2455 or 2457), the excess bandwidths will be reallocated to all terminals with unsatisfied bandwidth one by one again.

After adjusting the transmitting bandwidth of the terminal with unsatisfied requested bandwidth, every terminal's bandwidth compensation value can be further adjusted based on the maximum bandwidth and the transmitting bandwidth, as shown in FIG. 10.

Step 250 includes updating the bandwidth compensation value according to the overspent extent of the allocated transmitting bandwidth, which means the excess bandwidth will be added to the bandwidth compensation value and the bandwidth which should give but without giving will be subtracted from the bandwidth compensation value in the next transmitting cycle. Please refer to FIG. 14. In step 250, compare every terminal's last transmitting bandwidth and the maximum bandwidth (step 251). When the transmitting bandwidth is larger than the maximum bandwidth, the excess portion, which is the excess bandwidth obtained from the transmitting bandwidth subtracting the maximum bandwidth, will be added to the bandwidth compensation value (step 253) to update the bandwidth compensation value (step 257). On the other hand, if the requested bandwidth is smaller than the maximum bandwidth (i.e. the allocated transmitting bandwidth is smaller than the maximum bandwidth), the allocatable bandwidth (i.e. the un-used bandwidth obtained from the maximum bandwidth subtracting the transmitting bandwidth) or the allocated remaining requested bandwidth (i.e. the un-used bandwidth obtained from the requested bandwidth subtracting the transmitting bandwidth) will be subtracted from the bandwidth compensation value (step 255) to update the bandwidth compensation value (step 257). Similarly, every terminal's bandwidth compensation value can be updated one by one by repeatedly performing the above steps (i.e. steps 251, 253 and 257, or 253 and 257).

In order to prevent transmission delay from jiggering over, a maximum transmission bandwidth limitation will be previously set up, which means every terminal has one maximum transmission bandwidth limitation in one transmitting cycle. Therefore, before or after the initial bandwidth allocation (i.e. step 220), every terminal's requested bandwidth can be previously adjusted based on the maximum transmission bandwidth limitation for the terminal (step 260), as shown in FIGS. 15A and 15B.

Please refer to FIGS. 16A and 16B. The adjusting method in step 260 is to compare every terminal's maximum transmission bandwidth limitation and requested bandwidth one by one (step 261), and use the maximum transmission bandwidth limitation to replace the requested bandwidth being used in the following procedures when the requested bandwidth is larger than the maximum transmission bandwidth limitation (step 263). Otherwise, the original requested bandwidth will be sustained (i.e. doesn't use the maximum transmission bandwidth limitation to replace the requested bandwidth) (step 265). They are represented in the following formulas.

$\begin{array}{cc}{R}_j^{\prime }=\{\begin{array}{cc}{R}_j,& \mathrm{if}{R}_j\le {\mathrm{B\_bound}}_j\\ {\mathrm{B\_bound}}_j,& \mathrm{if}{R}_j>{\mathrm{B\_bound}}_j\end{array}& \left(11\right)\end{array}$

In the formula (11), R′_j is the modified requested bandwidth, R_j is the original requested bandwidth, and B_bound_j is the maximum transmission bandwidth limitation of the terminal. Here, the maximum transmission bandwidth limitation can be a bandwidth between the maximum bandwidth and the twice the maximum bandwidth.

According to the description above, in the transmitting bandwidth allocation method according to the disclosure, when a terminal requesting a bandwidth smaller or equal to the assured bandwidth, the bandwidth requested will be given. However if the bandwidth requested is larger than the assured bandwidth, the remaining bandwidth of the network will be first calculated after all the terminals have their assured bandwidth (or requested bandwidth), which will be further allocated based on a ratio of a value obtained from the maximum bandwidth of the terminal which has not finished the allocation subtracting the overspent bandwidth (i.e. bandwidth compensation value). In other words, the larger the maximum bandwidth the terminal has, the more bandwidth the terminal can get. The more overspent extent the terminal has, the less bandwidth the terminal can get. Because the maximum bandwidth is mostly set up based on the costumer contract, this method can accomplish the purpose of determining the remaining bandwidth allocation based on how important the client is and the excess use extent of the bandwidth so that the usable bandwidth of the network can be more fairly to be allocated to every terminal. Due to the set up of the maximum transmitting bandwidth limitation, the overspent bandwidth (i.e. bandwidth compensation value) will be limited in the maximum bandwidth. Therefore when the maximum bandwidth minus the cumulated overspent bandwidth (i.e. the bandwidth compensation value) is zero, this terminal will not be able to be included in the allocation of the remaining bandwidth so that a costumer's usable transmitting bandwidth can be effectively restricted, improved the fairness of the bandwidth allocation.

Next, focus on using the uploading order to improve the network transmitting bandwidth allocation, which adjusts the transmitting sequence mainly based on the data uploading amount of the terminal. This transmitting sequence is an order that the office terminal depends on to determine which uploading requested terminal to upload the data. Here, every terminal will deliver an uploading message to inform the office terminal before uploading the data to the office terminal through the network, where this uploading message includes a requested bandwidth to inform the office terminal the amount of the prepared uploading data. Please refer to FIG. 17. The office terminal will then obtain the requested bandwidths of all the terminals to which uploading messages are delivered (step 310). Next, sort the uploading order of the terminals to which uploading messages are delivered to get a transmitting sequence (step 320). Sequentially adjust the uploading order for every terminal in the transmitting sequence based on the size of the requested bandwidth (step 330), where the office terminal can then determine which terminal to upload data one by one.

The adjusting method in step 330 is to compare the requested bandwidths of the two terminals abutting each other in uploading order (step 331). If the requested bandwidth of a terminal with a lower uploading order is smaller than that of a terminal with a higher uploading order, the uploading orders of these two terminals will exchange (step 333). Otherwise the original uploading sequence will maintain the same (step 335) to get the modified transmitting sequence as shown in FIG. 18A. The above steps will repeat until all the terminals' uploading orders are adjusted.

For example, an embodiment is shown in FIG. 18B. Presume that there are j pieces of terminals to which uploading messages are delivered (i.e. the last uploading order is j). When comparing the requested bandwidths of a terminal with N−2 uploading order and a terminal with the N−1 uploading order (step 431) (where N≦j), and the requested bandwidth of the terminal with N−1 uploading order is smaller than that of the terminal with the N−2 uploading order, these two terminal's uploading order will switch (step 333). The uploading order of the terminal with the N−2 uploading order will change to the N−1 uploading order and the uploading order of the terminal with the N−1 uploading order will change to the N−2 uploading order. On the other hand, if the requested bandwidth of the terminal with N−1 uploading order is not smaller than that of the terminal with the N−2 uploading order, these two terminal's uploading order will not switch (step 335). Next, when the order has been switched, make sure whether the N+1 uploading order is the last uploading order or not (i.e. make sure whether N+1 equals to j or not) (step 437). If the N+1 uploading order is not the last uploading order (i.e. N+1≠j), continue to compare the requested bandwidth of a terminal with N uploading order and a terminal with the N+1 uploading order (step 441); otherwise (i.e. N+1=j), don't continue. If there is no switch between two orders, make sure whether the N uploading order is the last uploading order or not (i.e. make sure whether N equals to j or not) (step 439). If the N uploading order is not the last uploading order (i.e. N≠j), continuously compare the requested bandwidth of a terminal with N−1 uploading order and a terminal with the N uploading order (step 443); otherwise (i.e. N=j), don't continue.

The step of confirming whether the N+1 uploading order is the last uploading order or not also (i.e. confirm whether N+1 equals j or not) can be performed after step 333 or step 335 (step 437). If the N+1 uploading order is not the last uploading order (i.e. N+1≠j), continue to compare the requested bandwidth of a terminal with N uploading order and a terminal with the N+1 uploading order (step 439); otherwise (i.e. N+1=j), don't continue, as shown in FIG. 18C.

Only one modification of the transmitting sequence is described. However, according to the embodiment, the step 330 can be repeatedly performed to make every terminal's uploading order changed and become a better sequence by the proposed uploading data amount. However, in order to prevent the uploading order of each terminal from changing too much, a predetermined value can be set up previously. Also, after a modified transmitting order is received (i.e. step 330), make sure whether the modified transmitting order and the original transmitting order are the same or not (step 350). If they are not the same, cumulate the number of changing (step 360), and confirm whether the cumulated number of changing reach the predetermined value or not (step 370). If the cumulated number of changing does not reach the predetermined value, go back to the step 330 to adjust the transmitting sequence based on the requested bandwidth again. On the other hand, if the cumulated number of changing does reach the predetermined value, stop adjusting the transmitting sequence. The office terminal will then use the last adjusted transmitting sequence to determine which terminal should upload data, as shown in FIG. 19.

In summary, the disclosed exemplary embodiments provides a method for allocating bandwidth of a network which is capable of adjusting the transmitting sequence based on the uploading data amount of the terminal. The disclosure further provides a method for allocating bandwidth of a network which is capable of adjusting the predicting bandwidth allocation ratio based on the loading extent of the network for effectively reducing the average transmission delay. The bandwidth efficiency, the fairness of the bandwidth allocation and the jittering of transmission delay can be improved.

The embodiments mentioned in this specification can be arbitrarily combined or used alone when applied in a network to improve the efficiency of data uploading.

While several embodiments have been set forth for the purpose of disclosure, modifications of the disclosed embodiments of the disclosure as well as other embodiments thereof may occur to those skilled in the art. Accordingly, the appended claims are intended to cover all embodiments, which do not depart from the spirit and scope of the invention.

Claims

1. A method of allocating a bandwidth of a network, the network including an office terminal and a plurality of peripheral terminals that communicate with the office terminal by time division multiplexing during a sequence of transmitting cycles, the method comprising:

receiving a plurality of requested bandwidths from a plurality of uploading messages delivered from the peripheral terminals;

sorting an uploading order of the peripheral terminals based on a size of the requested bandwidths of uploading messages to obtain a transmitting sequence; and

adjusting the uploading order of each of the peripheral terminals in the transmitting sequence based on the size of the requested bandwidth to obtain a modified transmitting sequence, the adjusting step including; comparing the size of the requested bandwidths of the two peripheral terminals abutting to each other in uploading order in turn; switching the uploading orders of these two peripheral terminals to obtain the modified transmitting sequence, if the size of the requested bandwidth of the peripheral terminal with a lower uploading order is smaller than the size of the requested bandwidth of the peripheral terminal with a higher uploading orders of these two terminals to obtain the modified transmitting order; and sorting a modified uploading order of the peripheral terminals based on the modified transmitting sequence; and

identifying the modified transmitting sequence;

cumulating a number of changes in the sequence and confirming the number of changes based on a predetermined value if the modified transmitting order changes, after adjusting the uploading order; and

adjusting the uploading order of the peripheral terminal if the number of changing doesn't reach the predetermined value.

2. The method of claim 1, wherein the step of the identifying the modified transmitting sequence comprises:

comparing the transmitting sequences before and after the modification to determine if the transmitting sequence after modification is not the same with the transmitting sequence before modification.